The number of targeted therapies available for the treatment of different cancers increases annually. However, for many of these agents, we do not have effective predictive biomarkers. The result is that patients who might benefit for certain therapies are not receiving them, and patients receive therapies from which they receive no benefit. In many cases, it is likely that an individual drug might be effective in only a small subset of a particular cancer, so that without a validated predictive biomarker, introduction to that cancer is impossible. W propose a novel and generalizable functional predictive biomarker approach, based on measuring death signaling early in the response to targeted therapies. We hypothesize that if we could identify those cancer cells that induce the strongest activation ofthe mitochondrial apoptotic pathway at early time points in response to a drug, we could thus identify those tumors that are most likely to respond in vivo. We are greatly assisted in this approach by our development of a novel tool, BH3 profiling, which can measure a cell's proximity to the threshold of apoptosis. Our approach is to treat cells with drugs, measure at short time points the extent to which apoptotic signaling is forcing cells toward the threshold of apoptosis, a method we call Dynamic BH3 Profiling. As we will show, we are now adept at making such measurements. Moreover, we already have exciting preliminary data that suggests that those cells that are pushed the most are also those that are destined to choose death as a cell fate. We propose to develop Dynamic BH3 profling first in AML cell lines, and then in murine AML primagrafts. Our output will be a set of drugs for which we have validated Dynamic BH3 profiling as a predicitve biomarker in vivo. These drugs might be useful as single agents, or as agents to increase priming to enhance sensitiity to conventional, curative agents. Acute myelogenous leukemia (AML) is an attractive disease in which to test our ideas because of our established expertise, established relationships, and access to clinical samples. Dynamic BH3 profling has the potential to expand the pharmacopeia and personalize therapeutic decision making for the AML patient.